State of the Global Semiconductor Supply Chain Report

For the past 60 years, CSCMP has worked to develop and share an understanding of just how critical the global supply chain is to the functioning of the global economy. Logistics, warehousing, distribution, and transportation have been and remain key supply chain activities within this mission. Today, it can be argued that the global economy IS the global supply chain. Yet many aspects of this essential trade fabric are opaque or invisible to those that depend on it. Over these six decades, much of the world has become comfortable with “globalization” and with the associated opportunity for labor arbitrage to reduce the cost of both sourcing raw materials and components, and outsourcing manufacturing and assembly. This evolution of global trade has concentrated many critical “supply chain” elements into a small number of locations, primarily in Asia. As we shall see in what follows, this concentration has also determined patterns in the movement of raw materials, components and finished goods that define global supply chains.

It can be argued that the pursuit of efficiency in operations and the use of capital efficiently has led to additional concentrations all through the supply chain. While industry concentration may have led to more aggregate system capacity, it also tends to lead to less “diversity” and flexibility – and thus disruptions will tend to have a bigger impact.

In parallel, supply chains have also become more complex and dispersed. The “end to end” chain that converts raw materials into finished goods now typically has dozens, often hundreds of “links” as raw materials are mined and then refined to become components, which become subassemblies which become finished products. Much of this multi-step “dispersed complexity” has been made possible by low labor costs, cheap transportation, low cost of capital and a reasonable global security environment.

What this means for semiconductor manufacturing
Semiconductor manufacturing demonstrates globalization, industry concentration and the rest of these factors. Manufacturing is concentrated in two large integrated design and manufacturing businesses (IDMs -- Intel and Samsung) that design and manufacture their own products, and three large “foundries” (TSMC, Global Foundries and, also, Samsung) who manufacture semiconductors designed by their customers. Outside of this concentrated capacity are dozens of other manufacturers who make smaller quantities of older designs or specialized and generally less sophisticated semiconductor products. As an extreme example of concentration, almost all the most complex logic devices are manufactured by one company (TSMC) in one location (the island of Taiwan). Almost all advanced semiconductor memory is manufactured by just three companies (Samsung and SK Hynix, based in South Korea, and Micron, based in the US).

From an inbound supply chain perspective, several thousand companies, spread around the globe, support these manufacturing processes, but even within this extensive community, concentration and complexity exists. One company, ASML, which is based in the Netherlands, manufactures the machines needed to make the most advanced semiconductors. The specialized optical equipment needed by ASML comes from one Swiss/German Company, Zeiss, which in turn depends on local specialized precision manufacturing sub-contractors for the tools that measure the quality of its optical products. The purified silicon needed to grow the crystals from which “wafers” are sliced come almost entirely from one Japanese company, Tokyo Electron, with its own network of specialized subcontractors. The software that designs complex semiconductors, manages the complex manufacturing processes, and allows finished semiconductors to be assembled and tested comes largely from a handful of US companies. Critical raw materials are sourced in only a few places globally and refined in even fewer locations.

Finished semiconductors are often shipped to specialized locations for assembly, test, and packaging and then on to customers who use them in sub-assemblies in products such as motherboards or disk drives, and finished goods like computers and phones and TVs. Increasingly just about everything that uses electricity and aspires to be “smart” or connected requires multiple semiconductors. By the time a piece of “raw” silicon makes its way into a finished product it can have traversed as many as 40 different locations in as many countries and will still have one final journey to make – to the market in which it will finally be sold to a consumer. Over the past 70 years, supply chains have adapted to make these multiple movements cheap, efficient, and secure.

Looked at in retrospect, capital efficiency has led to both specialization where parts and products are made by the most efficient provider and user of capital, and concentration. There are few specialized manufacturers because the capital investment and the level of skill and experience required creates significant barriers to entry.

The global semiconductor supply chain
The story of the growth of the semiconductor industry is a story about the development of supply chains. Strong supply chains enabled the combination of computing, data storage and connectivity that has transformed our modern lives. The semiconductor sector is an example of how a set of complex international supply chains has facilitated a precision manufacturing process, involving the most complex machines, the purest materials, and the most expensive and specialized manufacturing techniques that humans have ever undertaken. These manufacturing processes have become essential to the entire global economy, yet they are poorly understood by most people who use the many products and services they make possible. Even beyond the complexities of device manufacturing, few people understand the complex supply chains that make that manufacturing possible. As we saw from the pandemic, without a better understanding of how the participants in the global semiconductor ecosystem fit together, it is difficult to assess where there are risks of disruption, and to design effective national policies that can address these risks. These issues apply across a broad spectrum from the availability of low cost finished consumer products and importantly, to national security.

Where do we go from here?
The post-pandemic realization that these supply chains are fragile and that concentration in key areas of supply and manufacturing exposes nations to national and economic security threats has prompted policy makers to look at how to both harden existing trade arrangements and redesign supply chains to better protect national concerns. The CHIPS and Science Act in the U.S. and similar initiatives in the European Union and the UK all seek to encourage the development of advanced semiconductor manufacturing capacity locally within their borders, or at least within their national control. However, the level of funding provided is small and is only $52.7 billion in the U.S. over a period of 5 years. These monies are focused primarily on developing industrial policy via incentives to create or expand “Lab to Fab” research and development and improved manufacturing processes. The Chips Act only incidentally touches on the other necessary capabilities such as raw material supply, reliable power and clean water, and assembly, test, and packaging (ATP) capacity. It does not focus much on the consequences for the network of associated supply chains.

And, as is often the case, there have already been unanticipated consequences, as the market value of semiconductor companies based in the Peoples’ Republic of China (PRC) has climbed, despite the potential impact of US-led trade restrictions. China is the largest consumer of semiconductors of all types because it is also the largest global assembler of finished products. Additional responses from the PRC have included restrictions on the export of essential raw materials and economic retaliation against US businesses that operate in or have customers in China. The end of this is far from clear and is likely to remain uncertain. There are many factors and interests involved in the attempt to establish a new equilibrium.

In a global economy increasingly dependent on the availability of semiconductors, it’s not enough to focus only on the need to promote and control the necessary fundamental research and the transfer of research results to manufacturing at scale. We need to create new public/private partnerships to ensure that the end-to-end processes needed to support resilient semiconductor manufacturing can function successfully. It is not possible for the government nor the semiconductor industry to do this alone.

We need to focus more on the consequences for the design and operation of reliable and resilient supply chains without a significant increase in the costs to move raw materials, manufactured devices, and finished goods. Supply chains matter. The trade fabric we have built up in the past 60 years may well need to change radically to meet the demands of the next decades.

At CSCMP we believe that we are in for an extended period of perhaps a decade or more of change in how these critically important supply chains are designed and executed. CSCMP will continue to monitor how these issues develop, how decisions being made today impact the outcomes we need, and where changes in strategy might become essential for success.